linuxcnc trajectory planner
26 Oct 2024 09:19 #313243
by Aciera
Replied by Aciera on topic linuxcnc trajectory planner
Yes that seems to work.
Note: Running './installer' ends with an error when trying to open a config in the end. But opening a new terminal then running 'source scripts/rip-environment' fixes it.
Note: Running './installer' ends with an error when trying to open a config in the end. But opening a new terminal then running 'source scripts/rip-environment' fixes it.
The following user(s) said Thank You: Grotius
Please Log in or Create an account to join the conversation.
28 Oct 2024 21:31 #313388
by Grotius
Replied by Grotius on topic linuxcnc trajectory planner
Hi Arciera,
the ./installer script seems to fail every time.
But ./build_lcnc && ./start_lcnc works fine. Maybe the problem is in ./deps_lcnc.
The custom trajectory planner component "tpmod.so" has already some logic inside now.
It is succesfully linked to multiple libraries. link.
It can load the gcode file using "tpAddline" & "tpAddCircle". "tpAddRigidTab" is also used, but later on we have
to talk about implementation properties like synchronised motion for tapping, wich are unknown for me now what to do.
This all then is stored into the vector library.
During loading it add's also empty fillet segments to the vector library and creates the fillets using the filletizer library based
on the given G64 values.
In the same flow the motionizer library calculates optimal trajectory velocity's using curvature extrema's, vm's, lenght's.
This optimalisation is done every gcode load cycle, as i don't know how to do it else. It's a fast algo, so i think it
Then step by step we make progress to the code.
We can decide if we use a max buffer size for our vector library. For example buffer 500 segment's max.
This is more or less how lcnc it does at the moment. We can also decide to make a ringbuffer. A buffer with
250 segments in the past and 250 segments in the future.
the ./installer script seems to fail every time.
But ./build_lcnc && ./start_lcnc works fine. Maybe the problem is in ./deps_lcnc.
The custom trajectory planner component "tpmod.so" has already some logic inside now.
It is succesfully linked to multiple libraries. link.
It can load the gcode file using "tpAddline" & "tpAddCircle". "tpAddRigidTab" is also used, but later on we have
to talk about implementation properties like synchronised motion for tapping, wich are unknown for me now what to do.
This all then is stored into the vector library.
During loading it add's also empty fillet segments to the vector library and creates the fillets using the filletizer library based
on the given G64 values.
In the same flow the motionizer library calculates optimal trajectory velocity's using curvature extrema's, vm's, lenght's.
This optimalisation is done every gcode load cycle, as i don't know how to do it else. It's a fast algo, so i think it
Then step by step we make progress to the code.
We can decide if we use a max buffer size for our vector library. For example buffer 500 segment's max.
This is more or less how lcnc it does at the moment. We can also decide to make a ringbuffer. A buffer with
250 segments in the past and 250 segments in the future.
Please Log in or Create an account to join the conversation.
02 Nov 2024 20:43 #313678
by akb1212
As one of the guys who plan to use LCNC for a wire-EDM machine I would hope this is the chosen path. After all it has been implemented and is working flawlessly as it is now. And it would be a great loss if a feature like that is removed.
At least make it possible to implement for those who need it. If it is possible to set it as an option if needed this is enough.
If not implementing it now means it will be difficult to implement at a later stage I say please dont leave it out at this stage.
And thanks for the great effort you are putting in to this Grotius!
Replied by akb1212 on topic linuxcnc trajectory planner
We can also decide to make a ringbuffer. A buffer with 250 segments in the past and 250 segments in the future.
As one of the guys who plan to use LCNC for a wire-EDM machine I would hope this is the chosen path. After all it has been implemented and is working flawlessly as it is now. And it would be a great loss if a feature like that is removed.
At least make it possible to implement for those who need it. If it is possible to set it as an option if needed this is enough.
If not implementing it now means it will be difficult to implement at a later stage I say please dont leave it out at this stage.
And thanks for the great effort you are putting in to this Grotius!
The following user(s) said Thank You: Grotius
Please Log in or Create an account to join the conversation.
05 Nov 2024 18:58 #313843
by Grotius
Replied by Grotius on topic linuxcnc trajectory planner
@Akb1212,
No worries. For wire edm you could set unlimited program history.
No worries. For wire edm you could set unlimited program history.
Please Log in or Create an account to join the conversation.
- tiagounderground
- Offline
- New Member
Less
More
- Posts: 8
- Thank you received: 6
20 Nov 2024 18:45 #314934
by tiagounderground
Replied by tiagounderground on topic linuxcnc trajectory planner
hi
will this version from your repository control an real machine with ethercat servos?
it already works with linuxcnc 2.9.3, its a cutting machine with XYA axis, A is tangential to the path. today it has very jerky moves as linuxcnc doesnt blend rotary axis. i would like to test the blend features you developed. in the previous videos there was an tool direction that i could use to control the tangencial axis, is this avaliable in this version?
will this version from your repository control an real machine with ethercat servos?
it already works with linuxcnc 2.9.3, its a cutting machine with XYA axis, A is tangential to the path. today it has very jerky moves as linuxcnc doesnt blend rotary axis. i would like to test the blend features you developed. in the previous videos there was an tool direction that i could use to control the tangencial axis, is this avaliable in this version?
Please Log in or Create an account to join the conversation.
Time to create page: 0.089 seconds